A Differentiable Integer Linear Programming Solver for Explanation-Based Natural Language Inference

arXiv:2404.02625v1 Announce Type: cross
Abstract: Integer Linear Programming (ILP) has been proposed as a formalism for encoding precise structural and semantic constraints for Natural Language Inference (NLI). However, traditional ILP frameworks are non-differentiable, posing critical challenges for the integration of continuous language representations based on deep learning. In this paper, we introduce a novel approach, named Diff-Comb Explainer, a neuro-symbolic architecture for explanation-based NLI based on Differentiable BlackBox Combinatorial Solvers (DBCS). Differently from existing neuro-symbolic solvers, Diff-Comb Explainer does not …

abstract arxiv challenges constraints continuous cs.ai cs.cl cs.lg deep learning differentiable encoding frameworks however inference integration language linear natural natural language programming semantic solver type